Health Policy

MedPAC Got It Wrong (pt 1)


This is the first part of former Kaiser Permanente CEO George Halvorson’s critique of Medpac’s new analysis of Medicare Advantage. The rest will be published on THCB later this week. Eventually I’ll be doing a summary article about all the back and forth about what Medicare Advantage really costs!-Matthew Holt

MedPac just did their annual report on Medicare Advantage (MA) and they were extremely wrong on several key points.

The MedPac staff has a long tradition of being critical of MA, and they also, unfortunately, have a long tradition of being inaccurate, misleading, and consistently negative on some key points for no explicable or easily understood reason.

They achieved a new low this year by spending more than 20 pages of the report warning us all in detail about the upcoming cash flow distortions and coding abuses that they say are coming from a risk adjustment model and system that actually no longer exists in 2022 as a functioning system for our Medicare program — and they are also continued their distortion about Medicare overpayment of the plans by running an artificial cost number that functions only to deceive and not to inform and by using what is essentially a fake news number several times in the report.

Coding and Risk Adjustment

CMS has now officially canceled and retired the CMS Hierarchical Conditions Categories Risk Adjustment Model that has been used for almost two decades to calculate risk for plans. It is dead and completely gone for 2022 — and MedPac explained bitterly for more than 20 pages why it was a damaging approach and they somehow did not mention that it was now gone.

CMS has some very good thinking people who brilliantly took that whole set of coding linked issues off the table by making the system that was being potentially abused simply disappear.

MedPac wrote more than 20 pages in this year’s official report about MA complaining about that exact process and system and they didn’t mention that it was gone or explain why it was important to not have that data flow create the risk level information that we will now be using to get diagnostic information into the system.

The new approach for determining patient risk levels is fraud proof. There is no way to put wrong data into the information flow that they are now going to use to see and determine which patients are diabetic and which have heart disease or who has drug abuse issues for the risk discernment processes.

The impact on low income Medicare patients & union members

MedPac also had a major content deficit in their report and managed to leave the most important aspects of the work being done now by the plans to help offset some of the damage done to too many Americans who have been damaged by social determinants of health issues for far too long in their lives. MedPac also completely failed to report and discuss the important reality of the fact that we have now reached the point where two-thirds of our lowest income Medicare beneficiaries are all voluntarily in the MA plans.

They also left out of their report the fact that a significant number of union trust funds and a significant number of employer retirement programs that had made significant promises of retirement health care benefits to their retirees over the past decades are actually having those commitments kept, met, and even enhanced with the relatively new employer-sponsored MA plans that work directly with employer settings.

Five million people who might have had their retirement health care programs bankrupt, underfunded, or at serious risk have found a very strong safety net in the MA program — and MedPac does not think that development was important to understand and probably celebrate.

Anyone looking at the future politics and funding of the MA program will find both that overwhelming support for MA from our lowest income people and from our most well-connected employer retirement funds to be good and important to understand.

MedPac missed every bit of that agenda and set of accomplishments in this year’s report.

MA vs FFS during Covid and beyond

They also left Covid out of the report as a care-related issue and the clear difference between the two programs.

They failed completely to explain the huge differences in the care reality, the personal comfort levels, and the direct care availability for Medicare beneficiaries. Care and treatment in the MA plans were significantly better on all of the Covid-related issues versus that for those people who were abandoned, isolated, and completely unsupported by the total non-response from fee-for-service Medicare.

MedPac should have used this report to explain what we have learned in that process — but they did not have any portion of the report that related to the impact of either program on actual people.

Leaving the impact of the program on people completely out of the report is not a new trajectory for that MedPac report. They do not see the impact of either program on actual people to be part of their role in advising Congress about Medicare issues.

In fact, they actually tend to complain every year that the vision benefits, hearing benefits, and dental benefits that are made available to everyone and that are appreciated most deeply by our lower income people and that happen only through the plans should not exist as benefits. The MedPac team usually says with a negative perspective that people’s expectations go up if they have eyeglasses and dental work and if they are able to hear and MedPac explicitly does not want to raise expectations of anyone on Medicare.

MedPac is actually consistently, explicitly and directly opposed to people having higher expectations and to better benefits for Medicare in those areas. They believe — incorrectly — that those benefits increase the overall cost of Medicare rather than actually use the Medicare cash flow more consistently and more effectively.

MedPac misses how much actually gets paid

They don’t understand, and they don’t believe, that the MA plans fund those additional benefits by making much better use of the Medicare dollar. That isn’t additional Medicare money. It results from the plans having costs that are below the average costs for fee-for-service Medicare in each area and it is possible for the plans to do so because fee-for-service Medicare buys care badly and delivers care poorly far too often with their much less competent use of the same dollar.

It actually costs less to buy care with MA . The plans have a process where they bid on what their monthly payment will be each year and the average bid is currently less than 85 percent of the average cost of paying for fee-for-service Medicare in each county.

They collect fewer Medicare dollars and they use them far better than those same dollars are used in the traditional payment approach that buys everything by the piece and literally does nothing in any intelligent way to get better value and higher leverage for the money they spend.

We know MA costs less because the bidding process is extremely transparent and open and it openly spends less to get more in plain sight.

The bids are actual bids and they result in actual payments. The MedPac sophists who argue as politicized theorists that the money that seems to be spent isn’t the actual cost because we need to revalue the dollars with a 9 percent judgment call to see what we are really spending because they are sure that a 9 percent inflation factor needs to be in the calculation even if it isn’t in the cash flow.

It’s a little like the old Marx movie that has MedPac saying: Who are you going to believe, me or your own eyes?

Believe your own eyes.

The bids in about a quarter of the counties fall below 80 percent of the average fee-for-service cost for each county — and stay there each year. When we look at how much we pay for each member, the actual dollars per member that is spent follow the bids — and the actual dollars go down compared to fee-for-service averages in almost all of those areas.

We don’t need to look at projections or estimates or theories on the cost of MA to figure out whether we are paying more or less. We can look at how much we actually pay and add up the dollars.

When plans bid 5 percent less and are paid 5 percent less than we can think of their costs as being 5 percent less.

MedPac does a report every year that contends that the actual amount we pay isn’t the important number for us to understand and we need to do some other calculations to see what the relative numbers actually are.

One report they do every year that distorts the relative cost of MA compared to fee-for-service Medicare looks at the actual care given to MA members and then prices that care using the Medicare fee schedule.

They look at MA enrollees and they look at their exact pattern of care and then they price that care using the Medicare fee-for-service fee schedule. They then compare that number with the capitation paid to the plans for those people and show that the numbers are consistently very different every year.

Using that model — and not adjusting it for the actual change in patterns of care — it generally looks like it costs about 4 percent more to pay the capitation than it costs to buy care for those people by the piece.

They love doing that report. It’s in this year’s annual report three times. The reality is that the care is far better for the MA plans and that’s the pattern of better care that gets priced by that process. The relevant numbers are all consistent and obvious.

MA has about 35 percent fewer emergency room days. That’s true every year. MA has programs to reduce the crisis level congestive heart failure events and that tends to result in 40 percent fewer hospital days for those patients. MA plans have many more eVisits. Fee-for-service Medicare does not pay for eVisits.

When MedPac does that cost comparison every year, they don’t have those emergency room visits in their care pattern, and that means that it looks like fee-for-service Medicare would have spent less on those exact patients.

They actually call that the overpayment report — and it typically shows that the Medicare would have spent about 4 percent less on those patients if the only care-related cash flow in their lives was the benefits that were included in the limited and standard traditional Medicare fee schedule and purchased by the piece for those exact patients.

It’s important for them to use the exact patients for that calculation, because when you use any other measurement, MA costs less. The capitation is lower than the average cost of care in all of those counties. That is extremely visible as an expense.

The plans each look at the average cost of fee-for-service Medicare in each county, and they get a chance to bid each year against those average costs and they always bid much less. In the highest cost counties for fee-for-service, even this year’s MedPac report shows that many MA plans are bidding 40 percent less.

The actual capitation that is paid to the plans each month is legitimately based on the actual average cost of care for fee-for-service Medicare in every county. That is a good, useful and important number and when we are looking at how much we are spending on each program, that average number is a very legitimate benchmark for figuring out if we are getting a good deal financially for the program.

They actually label that the benchmark process and it creates the benchmark prices that are used for enrollees in each county.

The capitation money the plans are paid each month is used by the plans to provide care, and the plans can be very flexible in how they use that money. The plans have to provide the basic set of Medicare benefits and they also need to do some defined and expected levels of quality and service reporting for that money, but they can use it in very flexible ways to provide care to their members.

They can use that money to provide additional levels of care and they can also use any and all legal caregivers available to them to serve their members in the process. A MA plan might send a nurse into a home to help with a diabetic patient or to help avert a potential congestive heart failure crisis, and they can pay that nurse to provide that care from the capitation paid to the plans. Fee-for-service Medicare does not pay for those nurses or for those sites of care — and fee-for-service Medicare would even accuse, charge and possibly indict that nurse for Medicare billing fraud if the nurse performed those same services and then sent a bill to Medicare for their time and efforts.

Quality Reporting in MA vs FFS

So MA patients tend to have both better care and more flexible care at several levels compared to fee-for-service Medicare patients. The plans also have quality reporting and expectations that do not exist at any level today in fee-for-service Medicare and that makes the MedPac comment in the report about the quality differences being completely non-discernable are discernably wrong.

We know from the current five-star reporting system that MA plans increased the percentage of cardiovascular patients who were receiving and using their statins from 80 percent to 83 percent even under Covid situations last year — and we know that the diabetics in MA plans lost ground on having their blood sugar controlled and the MA plans, on average, had that number with blood sugar controlled drop from 80 percent to 76 percent of the diabetic patients under the Covid year.

So, even though the plans did far better than fee-for-service Medicare in responding to the immediate care needs of their Covid patients, we know that they lost some ground on some levels of care during that time.

Fee-for-service Medicare, by contrast, actually does not have one single quality measure.


Fee-for-service Medicare simply does not do any quality improvement programs. They also don’t set any quality goals, and fee-for-service Medicare does not support or pay for team care or for medical information pooling and linkages for their patients for use in care quality activities for their caregivers.

The MedPac staff wrote again in their report about the status of MA that they could not comment on the quality of care in fee-for-service Medicare compared to MA because they have no data for those fee-based patients, which means that it is impossible and wrong for MedPac to comment on the comparative quality of care. They believe and they say that clear deficit in quality data that exists for fee-for-service care sites and those caregivers should not cause anyone to conclude anything about the relative quality of the two programs, or should be a reflection in any way on any part of the fee-for-service Medicare, because that would be “unfair to fee-for-service Medicare” to point that data issue out in any public way.

That avoidance of the quality issue is just wrong.